Efficacy and safety of memantine as adjunctive treatment for manic episodes in bipolar disorder (BD): a randomized, double-blind, placebo-controlled clinical trial

Abstract

Background

Bipolar disorder (BD) is a chronic and recurrent psychiatric illness, and acute manic episodes are associated with marked functional impairment. Mood stabilizers and antipsychotics remain the mainstay of treatment. Dysregulation of glutamatergic neurotransmission and N-methyl-D-aspartate (NMDA) receptor signaling has been implicated in bipolar disorder (BD), providing a rationale for evaluating memantine as an adjunctive treatment for manic episodes in bipolar I disorder (BD-I).

Methods

In this 6-week randomized, double-blind, placebo-controlled trial, adults aged 18–65 years with DSM-5 BD-I and a current manic episode were allocated 1:1 to adjunctive memantine or placebo, alongside lithium and olanzapine. Memantine was titrated from 5 mg/day to 20 mg/day. Manic symptoms were assessed using the Young Mania Rating Scale (YMRS), with secondary outcomes including Clinical Global Impression (CGI) scales and peripheral inflammatory markers.

Results

Sixty-three participants were randomized (memantine n = 31; placebo n = 32). Baseline YMRS scores were comparable between groups. Mean YMRS scores decreased significantly in both groups over time; however, no between-group differences were observed at week 2 (P = 0.674), week 4 (P = 0.778), or week 6 (P = 0.466). At week 6, the magnitude of YMRS reduction did not differ significantly between the memantine and placebo groups. CGI-Severity, CGI-Improvement, and CGI-Efficacy Index scores were also similar between groups at all assessment points (all P > 0.05). Changes in white blood cell count, neutrophil count, lymphocyte count, and neutrophil-to-lymphocyte ratio at week 6 and week 18 showed no statistically significant differences between the memantine and placebo arms in either intention-to-treat or per-protocol analyses.

Conclusions

Adjunctive memantine did not provide additional symptomatic or biological benefit beyond lithium and olanzapine in the treatment of acute mania in BD-I. These findings do not support routine use of memantine for acute manic episodes, although further research may be warranted in selected patient subgroups or alternative clinical contexts.

Trial registration

Iranian Registry of Clinical Trials (IRCT), IRCT20250430065539N1, 30 April 2025.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12888-026-08398-y.

Introduction

According to the DSM-5-TR, bipolar disorders (BD) represent a spectrum of chronic mood conditions characterized by recurrent episodes of mood elevation, from hypomania to mania, commonly alternating with depressive episodes and often misdiagnosed at initial presentation. The spectrum includes bipolar I and II disorders, cyclothymic disorder, and other specified or unspecified BD [1]. In line with updates to psychiatric classification, ICD-11 replaced the ICD-10 term “bipolar affective disorders” with “bipolar disorders” and organized them under the broader category of “bipolar and related disorders,” closely aligning its terminology and structure with the DSM-5 [2]. Population-based surveys and meta-analyses report a global lifetime prevalence of bipolar disorder, based on threshold DSM diagnostic criteria, ranging from approximately 0.7% to 1.3%. When broader bipolar spectrum conditions are considered, including subthreshold presentations, prevalence estimates increase to nearly 2% to 3% in certain populations [3]. BD is linked to significant impairment of functions, diminished living standards, and high morbidity and mortality [4, 5]. Individuals with BD have a markedly reduced life expectancy of approximately 12 to 14 years compared with the general population, largely attributable to elevated rates of metabolic syndrome, diabetes mellitus, cigarette smoking, cardiovascular disease, and suicide, with an annual suicide rate of approximately 0.9% and a lifetime suicide risk of 15% to 20% [6–10]. In parallel, biomarker-oriented psychiatric research (including microRNA dysregulation associated with suicidality phenotypes) may ultimately improve mechanistic understanding and risk stratification relevant to mood disorders [11]. It makes it very clear that more effective and safer therapeutic approaches are urgently required, especially when dealing with acute mood episodes.

The treatment of BD, especially manic episodes, is a challenging situation even with the presence of pharmacological treatment. Standard treatment for BD relies on mood stabilizers such as lithium, valproate, and lamotrigine, along with atypical antipsychotics including quetiapine, aripiprazole, asenapine, lurasidone, and cariprazine, supported when appropriate by antidepressants and psychosocial approaches like cognitive-behavioral therapy and psychoeducation. However, many of these medications are associated with metabolic side effects, limited tolerability, and poor adherence, and more than half of patients struggle to remain on long-term treatment [6, 12]. Moreover, the antidepressant therapy in BD is disputable because some of the treatments may increase the mood fluctuation or trigger manic switching, particularly in BD I [13–15]. Such restrictions underscore the need to add adjunctive therapies with new mechanisms of action and safety profiles.

There is an increasing body of evidence implicating glutamatergic dysregulation, especially N-methyl-D-aspartate receptor (NMDAR) mediated signaling, in the pathophysiology of BD [16, 17]. Notably, established mood stabilizers such as lithium and valproate have been shown to modulate NMDAR activity, suggesting the therapeutic relevance of this pathway [18, 19]. Memantine is a low-affinity, uncompetitive NMDAR antagonist that selectively blocks pathological extrasynaptic NMDAR activation while preserving normal synaptic transmission, thereby reducing excitotoxic neuronal damage [20, 21]. Preclinical and clinical evidence indicate that memantine could have antimanic and mood-stabilizing properties that involve inhibition of dopamine receptor sensitization and reduction of glutamate-mediated excitotoxicity linked to mania [22]. Although memantine has a well-established safety profile and has been in clinical use since 1982 [23, 24], evidence supporting its efficacy in BD, particularly in acute manic episodes, remains limited and inconclusive.

To date, well-designed randomized controlled trials (RCTs) assessing memantine for acute mania are scarce, and evidence regarding its efficacy and safety remains inconclusive. Accordingly, the present interventional RCT aims to evaluate the efficacy and adverse events of memantine as an adjunctive treatment for manic episodes in patients with BD-I.

Methods

Study design

This double-blind, randomized, placebo-controlled clinical trial was designed to assess the efficacy and safety of memantine as an adjunctive therapy for acute manic episodes in BD-I. The study was conducted at Shiraz University of Medical Sciences and included hospitalized patients with acute mania who were receiving inpatient psychiatric care during the treatment period. Eligible patients were allocated to the memantine or placebo groups in a 1:1 ratio using block randomization. This randomized, double-blind, placebo-controlled clinical trial was registered in the Iranian Registry of Clinical Trials (IRCT) under the registration number IRCT20250430065539N1.

Participants

Adult patients aged 18–65 years who met the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria for BD-I and were experiencing an acute manic episode were eligible for inclusion in the trial; diagnosis was confirmed by clinical interview and symptom assessment consistent with DSM-5 standards. Consistent with prior randomized drug trials in bipolar mania, participants were required to be medically stable and able to comply with study procedures, and were recruited from both inpatient psychiatric units and outpatient settings. Individuals were excluded if they had current comorbid substance use disorders, significant uncontrolled medical or neurological conditions, suicide risk requiring crisis intervention, pregnancy or breastfeeding, or any contraindication to study medications, reflecting common exclusion criteria designed to protect participants and ensure interpretability of treatment effects in bipolar clinical research. Baseline insight was assessed during the psychiatric examination using the standard clinical description of insight described in Kaplan and Sadock’s Comprehensive Textbook of Psychiatry, and was rated on a 0–4 ordinal scale from complete denial of illness to true emotional insight.

Randomization

Randomization in this study was achieved using block randomization to ensure balanced allocation of participants across the two treatment groups: memantine and placebo. The block randomization method was employed to prevent any imbalances in group sizes at different stages of the study. A random number table was used to assign participants to one of two groups in a 1:1 ratio, ensuring that each group contained an equal number of participants at baseline.

Blinding

This study utilized a double-blind design, meaning that both the participants and the study personnel were unaware of the treatment allocation. The memantine and placebo pills were identical in appearance, size, and packaging, ensuring that neither the participants nor the researchers could distinguish between the active drug and the placebo. The randomization process was performed by a separate team of researchers who were not involved in any direct patient care or outcome assessment. To further ensure the integrity of the blinding process, the investigators conducting the assessments of the outcome measures, including YMRS, CGI scores, and adverse events (AEs), were blinded to the participant’s group allocation. Data collection was performed by a separate team that was blinded to the treatment assignment. Only the treating physician, who was responsible for managing the patient’s condition, was aware of the participant’s group assignment and the corresponding treatment. This design was implemented to reduce performance and detection biases and to ensure that the outcomes were assessed objectively.

Intervention

Participants in both groups received standard treatment with lithium and olanzapine, which are commonly used for managing manic episodes in BD-I and are supported by major bipolar disorder treatment studies [25–27]. The medication variables reported in Table 1 refer to medication use before study enrollment and do not indicate different background treatment during the trial. During the study period, all randomized participants received the same background pharmacological regimen with lithium and olanzapine; therefore, the only planned treatment difference between the two arms was adjunctive memantine versus matching placebo. The memantine group received memantine as an adjunctive treatment, with the following dosing schedule: 5 mg/day for the first 10 days, 10 mg/day for the next 10 days, and 20 mg/day for the remaining duration of the study. The placebo group received identical placebo pills that matched the appearance of memantine tablets but contained no active ingredient and were administered according to the same schedule. Treatment was administered daily over a 6-week period, and patients were monitored for both efficacy and AEs. Because all participants were inpatients, medication administration was directly supervised by ward nurses and recorded as part of routine inpatient care. Adherence was additionally checked through questioning family members/caregivers when applicable and by counting the remaining study tablets during follow-up assessments. All participants were hospitalized patients, and follow-up assessments were performed face-to-face at weeks 2, 4, and 6 by trained study assessors. At each visit, manic symptoms, CGI scores, AEs, and medication adherence were evaluated. Because participants were receiving inpatient psychiatric care, clinical monitoring was continuous during hospitalization, and family members/caregivers were also contacted when needed to support participation and verify adherence-related information.

Table 1.

Demographic and baseline characteristics of the participants

		Total N = 63	Placebo N = 32	Memantine N = 31	P-value
Age (years old)		36.19 ± 12.05	36.34 ± 13.3	36.03 ± 10.82	0.919 T
Children		0.97 ± 1.29	1.25 ± 1.52	0.68 ± 0.94	0.189 M
Number of admission		3.98 ± 3.44	4.47 ± 4.01	3.48 ± 2.72	0.596 M
Gender	Male	33 (52%)	17 (53%)	16 (52%)	0.904 C
	Female	30 (48%)	15 (47%)	15 (48%)
Residence	Rural	10 (16%)	6 (19%)	4 (13%)	0.732 F
	Urban	53 (84%)	26 (81%)	27 (87%)
Psychotic feature	No	13 (21%)	5 (16%)	8 (26%)	0.318 C
	Yes	50 (79%)	27 (84%)	23 (74%)
Educational level	Under diploma	23 (37%)	12 (38%)	11 (36%)	> 0.999 FF
	Diploma	31 (49%)	16 (50%)	15 (48%)
	University	9 (14%)	4 (13%)	5 (16%)
Past ECT history	No	40 (63%)	19 (59%)	21 (68%)	0.490 C
	Yes	23 (37%)	13 (41%)	10 (32%)
	Yes	48 (76%)	27 (84%)	21 (68%)
Lithium	No	39 (62%)	22 (69%)	17 (55%)	0.256 C
	Yes	24 (38%)	10 (31%)	14 (45%)
Benzodiazepine	No	37 (59%)	18 (56%)	19 (61%)	0.685 C
	Yes	26 (41%)	14 (44%)	12 (39%)
Substance history	No	24 (38%)	11 (34%)	13 (42%)	0.537 C
	Yes	39 (62%)	21 (66%)	18 (58%)
	Yes	16 (25%)	6 (19%)	10 (32%)
Insight	1	40 (63%)	22 (69%)	18 (58%)	0.479 FF
	2	15 (24%)	8 (25%)	7 (23%)
	3	7 (11%)	2 (6%)	5 (16%)
	4	1 (2%)	0 (0%)	1 (3%)

Values are mean ± standard deviation or number (percentage)

T: P-value of independent t-test, M: P-value of Mann-Whitney U test, C: P-value of Pearson Chi-Square test, and FF: P-value of Fisher-Freeman-Halton exact test

Outcomes

Young Mania Rating Scale (YMRS)

The YMRS was the primary efficacy outcome for this study. The YMRS is a clinician-rated scale designed to quantify the severity of manic symptoms based on a semi-structured interview and clinical observation over the preceding ~ 48 h. It includes 11 items with a total score range of 0–60; 7 items are scored 0–4 and 4 key items are double-weighted (0–8) to better capture core manic features (e.g., irritability, speech, thought content, disruptive/aggressive behavior). Higher scores indicate more severe mania, and the YMRS is widely used in acute mania trials because it is sensitive to clinical change and shows strong inter-rater reliability when raters are trained. For the Persian version, Iranian psychometric work supports acceptable internal consistency and validity; for example, one standardization study reported Cronbach’s α = 0.75 and concurrent validity r = 0.74 (correlation with a bipolar depression rating measure) [28].

Clinical Global Impression (CGI) scales

The Clinical Global Impressions (CGI) scales are brief clinician-rated global measures originally developed for psychopharmacology trials and remain widely used across psychiatric RCTs. CGI-Severity (CGI-S) rates current illness severity on a 7-point anchored scale from 1 = normal/not ill to 7 = among the most extremely ill; CGI-Improvement (CGI-I) rates change from baseline on a 7-point scale from 1 = very much improved to 7 = very much worse. The CGI-Efficacy Index (CGI-E) integrates perceived therapeutic benefit with the burden of adverse events/side effects using anchored categories in the original CGI materials and commonly used CGI forms [29].

Hematological markers

Exploratory inflammatory outcomes included longitudinal measurements of white blood cell (WBC) count, neutrophils, lymphocytes, and neutrophil‑to‑lymphocyte ratio (NLR) at weeks 0, 2, 4, 6, and 18. Blood samples were collected during face-to-face assessments. For the week-18 follow-up, participants were invited to attend an in-person visit specifically for clinical review and blood sampling; however, some week-18 samples were missing. All hematological tests were performed in the same laboratory using the same routine laboratory procedures to minimize measurement variability. NLR, calculated as the ratio of absolute neutrophil count to lymphocyte count, serves as a readily derived marker of systemic inflammation and has been investigated in mood disorders as a potential state marker associated with symptom severity. Associations between changes in these hematological measures and YMRS improvement were also explored.

Adverse events

Safety outcomes focused on the incidence of prespecified AEs, including constipation, dizziness, headache, confusion, fatigue, vomiting, and agitation, recorded at weeks 2, 4, and 6 and cumulatively over the treatment period. AEs were assessed face-to-face using a prespecified checklist. For each AE, severity was rated using four categories: “not present,” “mild,” “moderate,” and “severe,” scored as 0, 1, 2, and 3, respectively. For the main safety analysis, the presence of an AE was defined as any score greater than 0. Standardized AE reporting guided comparisons of AE rates between the memantine and placebo groups, enabling assessment of tolerability in addition to efficacy. The AE checklist has been provided in the Supplementary Materials.

Statistical analysis

Continuous variables were summarized as mean ± standard deviation (SD), and categorical variables were presented as number (percentage). The normality of continuous variables was assessed through Shapiro-Wilks test. Between-group comparisons at baseline were performed using the independent samples t-test for normally distributed variables and the Mann–Whitney U test for non-normally distributed variables. Categorical variables were compared using the Pearson chi-square test, Fisher’s exact test or the Fisher–Freeman–Halton exact test. Changes in biomarkers and clinical scales over time were calculated as the difference between baseline and follow-up measurements. Between-group comparisons of these changes were conducted using independent t-tests or Mann–Whitney U tests, depending on data distribution. Biomarker analyses were performed using both per-protocol (PP) and intention-to-treat (ITT) approaches, due to loss to follow-up. For week-18 biomarker analyses, available face-to-face blood samples were included in the PP analysis, and missing week-18 biomarker values were handled in the ITT analysis according to the prespecified missing-data approach. All statistical tests were conducted using IBM SPSS 24, and a P-value < 0.05 was considered statistically significant.

Results

A total of 173 individuals were assessed for eligibility, of whom 113 were excluded mainly due to not meeting inclusion criteria or declining participation. Sixty-three participants were randomized to memantine (n = 31) or placebo (n = 32). All randomized participants completed the face-to-face week-6 assessment. No participants discontinued treatment or were lost to follow-up. No participants discontinued treatment or were lost to follow-up, and all randomized patients were included in the week-6 analysis (Fig. 1). There were no statistically significant differences between the two groups regarding demographic characteristics, clinical variables, psychiatric history, medication use, or substance use (P > 0.05 for all comparisons). All characteristics were well balanced between the groups. All participants were hospitalized inpatients (Table 1).

Fig. 1.

Flow diagram of all included participants

No statistically significant differences were observed between the memantine and placebo groups in changes of inflammatory biomarkers, including WBC count, neutrophil count, NLR, and lymphocyte count, at either week 6 or week 18. This lack of significant difference was consistent across both per-protocol and intention-to-treat analyses (P > 0.05 for all comparisons). Although both groups showed numerical changes in biomarker levels over time, the magnitude of change did not differ significantly between treatment arms, suggesting that memantine did not have a measurable effect on these biomarkers compared with placebo during the study period (Table 2) (Supplementary Material, Figures S1-S4).

Table 2.

Between groups comparison of change in the biomarkers during the study

Biomarker	Group	Baseline vs. Week 6 N = 63		Baseline vs. Week 18 N = 52
		PP		PP		ITT
WBC	Memantine	-0.71 ± 1.13	P = 0.873	-1.06 ± 1.18	P = 0.226	-1.01 ±	P = 0.375
	Placebo	-0.67 ± 0.88		-0.66 ± 1.15		-0.66 ±
Neutrophil	Memantine	-4.74 ± 5.96	P = 0.227	-3.85 ± 6.61	P = 0.810	-3.05 ±	P = 0.845
	Placebo	-2.31 ± 10.84		-4.27 ± 6.02		-2.58 ±
NLR	Memantine	-0.50 ± 1.19	P = 0.768	-0.53 ± 1.34	P = 0.547	-0.44 ±	P = 0.670
	Placebo	-0.33 ± 0.60		-0.36 ± 0.50		-0.33 ±
Lymph	Memantine	5.00 ± 6.02	P = 0.837	4.96 ± 6.79	P = 0.803	4.45 ±	P = 0.586
	Placebo	5.31 ± 6.00		5.42 ± 6.45		5.46 ±

Values are mean ± standard deviation

PP: Per Protocol analysis, ITT: Intention to treat analysis

P-values are derived from independent t-test

Both the memantine and placebo groups demonstrated substantial reductions in YMRS scores from baseline to weeks 2, 4, and 6, indicating clinical improvement over time. The mean YMRS score decreased from 33.55 at baseline to 13.52, 9.23, and 5.00 at weeks 2, 4, and 6 in the memantine group, and from 31.59 at baseline to 13.25, 8.68, and 4.97 in the placebo group, respectively. However, there were no statistically significant differences between the two groups in the magnitude of YMRS score reduction at any assessed time point (P = 0.674 at week 2, P = 0.778 at week 4, and P = 0.466 at week 6). These findings suggest that improvement in manic symptoms was comparable between the memantine and placebo groups throughout the study (Table 3) (Supplementary Material, Figures S5).

Table 3.

Mean YMRS scores and between-group comparison of changes in YMRS during the study (N = 63)

Group	Baseline YMRS, mean ± SD	Week 2 mean score	Week 4 mean score	Week 6 mean score	Change to W2, mean ± SD	Change to W4, mean ± SD	Change to W6, mean ± SD
Memantine	33.55 ± 7.70	13.52	9.23	5.00	-20.03 ± 6.12	-24.32 ± 7.30	-28.55 ± 7.68
Placebo	31.59 ± 8.99	13.25	8.68	4.97	-18.34 ± 9.05	-22.91 ± 9.58	-26.62 ± 9.31

Values are shown as mean ± standard deviation for baseline and change-from-baseline values. Week 2, week 4, and week 6 YMRS score columns show mean follow-up scores

Change values were calculated as follow-up minus baseline. P-values for between-group comparison of change from baseline were derived from the Mann–Whitney U test: P = 0.674 at week 2, P = 0.778 at week 4, and P = 0.466 at week 6

Comparisons of Clinical Global Impression subscales (CGI-Severity, CGI-Efficacy, and CGI-Improvement) showed no statistically significant differences between the memantine and placebo groups at weeks 2, 4, or 6 (P > 0.05 for all comparisons). Both groups exhibited gradual improvement in CGI scores over time; however, the extent of improvement did not differ significantly between treatment arms. These results indicate that adjunctive memantine did not result in superior global clinical outcomes compared with placebo during the study period (Table 4) (Supplementary Material, Figures S6-8).

Table 4.

Between groups comparison of CGI subscales during the study (N = 63)

Biomarker	Group	Week 2		Week 4		Week 6
CGI-S	Memantine	4.00 ± 0.73	P = 0.314	3.19 ± 0.95	P = 0.244	2.32 ± 0.98	P = 0.303
	Placebo	3.84 ± 0.85		3.00 ± 0.84		2.03 ± 0.59
CGI-E	Memantine	5.61 ± 2.40	P = 0.520	4.55 ± 2.25	P = 0.344	1.77 ± 1.61	P = 0.259
	Placebo	6.06 ± 1.93		3.91 ± 2.12		2.28 ± 1.94
CGI-I	Memantine	2.32 ± 0.60	P = 0.383	2.06 ± 0.57	P = 0.433	1.42 ± 0.56	P = 0.845
	Placebo	2.22 ± 0.49		1.97 ± 0.40		1.37 ± 0.49

Values are mean ± standard deviation

P-values are derived from Maan-Whitney U test

Regarding safety outcomes, adverse events were systematically assessed at weeks 2, 4, and 6 using the prespecified AE checklist. Although severity scores were calculated for each reported event, all recorded adverse events were mild. No moderate or severe adverse events were reported in either the memantine or placebo group during the treatment period.

Discussion

In our study, although both groups experienced significant reductions in manic symptoms, adding memantine to standard therapy did not provide any additional benefit over placebo in alleviating mania or enhancing overall clinical outcomes in patients with BD-I. Inflammatory markers showed no significant between-group differences. Interest in memantine as a potential treatment for BD-I stems from accumulating evidence implicating NMDAR–mediated glutamatergic dysfunction in the disorder. Altered NMDA-mediated transmission may be partly driven by autoimmune mechanisms, as acute mania has been associated with elevated titers of anti-NMDAR antibodies [30, 31]. Mood stabilizers such as lithium and valproic acid are also known to modulate NMDAR activation, although this is not considered their primary mechanism of action [19, 32]. In the present trial, valproic acid was cited as part of the broader mechanistic literature linking established mood stabilizers with glutamatergic/NMDAR-related pathways; however, the study was not designed to compare lithium with valproate or to test different background mood stabilizers. Lithium and olanzapine were selected as the standardized background regimen because both are established treatments for acute mania, supported by systematic-review evidence and randomized clinical trials [25–27]. Using the same lithium-plus-olanzapine regimen in both arms minimized pharmacological variability and allowed the trial to specifically evaluate whether adjunctive memantine provided additional benefit beyond accepted antimanic treatment. Therefore, the substantial improvement observed in both groups may largely reflect the effect of standardized antimanic therapy and structured inpatient care, leaving limited room for memantine to demonstrate incremental efficacy. It has been suggested that lithium may exert part of its antimanic and mood-stabilizing effects through immunomodulatory mechanisms that reduce anti-NMDAR antibody production [30]. Genetic and postmortem studies further support the involvement of NMDAR dysfunction in BD. Variants in NMDAR–encoding genes have been associated with BD and related psychiatric conditions [33], with specific GRIN1 haplotypes strongly linked to BD across diagnostic subtypes [34]. Certain single nucleotide polymorphisms in NMDAR genes have been associated with markers of illness severity, such as frequency of manic hospitalizations [35]. Neuroimaging and postmortem studies have demonstrated reduced expression of NMDAR subunits and associated postsynaptic proteins in key brain regions implicated in mood regulation, including the hippocampus, dorsolateral prefrontal cortex, anterior cingulate cortex, and thalamus [36–41].

The absence of additional benefit from memantine in our study may be explained by its distinct pharmacodynamic properties. Memantine acts as an uncompetitive, low-affinity NMDA receptor antagonist, preferentially inhibiting excessive receptor activation while largely sparing physiological neurotransmission, with effects that depend on prior receptor activation [42]. During acute manic episodes, NMDA receptor overactivity may not be sufficiently prominent or central to symptom expression, limiting the clinical impact of NMDA modulation [42]. Memantine’s relatively low affinity for the NMDA receptor PCP-binding site (Ki = 740 nM), compared with high-affinity antagonists such as (+) MK-801 or phencyclidine (PCP), restricts the extent of receptor blockade achievable at therapeutic doses [43–45]. Its relatively fast off-rate allows rapid dissociation from the NMDA channel, preserving normal synaptic transmission [42, 44–47]. Although these features contribute to its favorable tolerability profile, they may also reduce its ability to suppress pathological glutamatergic signaling during acute mania. Additionally, memantine’s voltage-dependent binding preferentially targets receptors during pathological depolarization, while transient physiological NMDA activation remains largely unaffected, further limiting its clinical efficacy [44, 45, 47, 48]. Furthermore, because both treatment arms received lithium and olanzapine, potent antimanic agents, the substantial symptomatic improvement achieved with standard therapy may have produced a high-response background, leaving limited scope for memantine to demonstrate incremental efficacy.

Our findings are consistent with the existing evidence regarding the limited efficacy of memantine as an adjunctive treatment for core mood symptoms in BD. Similar to us, in a systematic review and meta-analysis by Zheng et al. in 2019, memantine did not outperform comparators in improving either depressive or manic symptoms in BD, suggesting a lack of robust mood-stabilizing efficacy across mood phases [49]. Nevertheless, previous individual studies have reported more favorable outcomes, particularly in specific populations or non-controlled designs. Omranifard et al. examined elderly patients with BD in the acute manic phase and found that although mania severity decreased significantly in both treatment arms, the reduction was greater in the memantine group at later time points, suggesting a potential long-term adjunctive effect when combined with sodium valproate [50]. Differences in patient age, treatment duration, and background pharmacotherapy may explain the discrepancy between those findings and the current study. Serra et al. reported that the addition of memantine reduced the duration and recurrence of mood episodes over a three-year follow-up period [51]; however, the absence of a control group may limit the interpretability of these results. In addition, Koukopoulos et al. observed substantial clinical improvement and stabilization in a sample of BD patients receiving memantine as add-on therapy, including among rapid cyclers [52]. Likewise, a pilot study in 35 patients with BD I by Keck Jr et al. reported improvement in manic symptoms across cohorts receiving memantine [53]. Although memantine has been studied in the manic phase of BD, Anand et al. assessed its adjunctive use in bipolar depression and found no significant advantage over placebo when combined with lamotrigine over eight weeks. Notably, a short-lived antidepressant effect was observed during the initial titration period, indicating a potential time- or dose-related effect that requires confirmation in larger controlled studies [54]. However, according to Krzystanek et al., despite its lack of efficacy in recurrent depression, memantine may be a safe and effective adjunctive treatment for bipolar depression [55]. Beyond mood symptoms, memantine has shown efficacy in reducing obsessive–compulsive symptoms in patients with BD during the manic phase, as demonstrated in a double-blind controlled trial by Sahraian et al. [56]. We did not find any significant difference in inflammatory biomarkers between the memantine and placebo groups. In a randomized, double-blind controlled trial, Lee et al. investigated the effects of adjunctive memantine on clinical symptoms and inflammatory cytokines in patients with BD II receiving valproic acid. Although depressive and manic symptoms improved over time in both treatment arms, memantine did not confer additional clinical benefit compared with placebo. Notably, the memantine group showed a significant reduction in plasma tumor necrosis factor-α levels, while no significant effects were observed for other cytokines [57]. The absence of significant differences in inflammatory markers between the treatment groups may be attributable to the non-specific nature of these measures. NLR is influenced by age, medications, chronic diseases, psychiatric conditions, cancer, anemia, and stress [58], while WBC counts can be elevated by infection, inflammation, allergic reactions, malignancy, or hereditary disorders [59]. These factors may have masked any subtle treatment-related effects of memantine.

Overall, evidence indicates that memantine’s effects in BD are domain-specific rather than broadly antimanic, with potential benefits limited to selected cases and not consistently replicated in randomized controlled trials. Consequently, memantine does not qualify as a mood stabilizer as monotherapy, though its adjunctive use in the treatment or prevention of BD may be considered [60]. Despite a biological rationale and favorable findings from open or uncontrolled studies, particularly in complex or rapid-cycling patients [51–53], memantine failed to demonstrate clinically meaningful antimanic efficacy under rigorous randomized conditions. Moreover, although memantine has been reported to modulate inflammatory pathways [57], no significant changes in peripheral inflammatory biomarkers were observed.

Strengths, limitations, and future directions

This study has several strengths. The randomized, placebo-controlled design enhances internal validity and minimizes biases. The baseline demographic and clinical characteristics were also well balanced across groups, which minimized the possibility of confounding variables. The use of validated and widely accepted clinical assessment tools, including the Young Mania Scale and Clinical Global Impression scales, ensured reliable and standardized evaluation of manic symptoms and global clinical outcomes. Several limitations should be acknowledged. A key limitation of our study is the relatively small sample size, and this might have diminished the statistical power to identify small to moderate treatment effects. In addition, the duration of follow-up for manic symptom assessment was relatively short, which may not fully capture delayed or longer-term effects of memantine. Inflammatory assessment was restricted to peripheral blood-based markers, which may not accurately reflect central neuroinflammatory processes implicated in BD.

Research in the future must be oriented towards bigger, sufficiently powered randomized controlled trials. Further follow-up studies should be considered to measure the possible delayed, neuroprotective, or maintenance-associated effects of memantine in the post-acute manic phase. Since there is emerging evidence of varying effectiveness of glutamatergic modulators in mood states, the impact of memantine in bipolar depression, cognitive impairment, and relapse prevention in the long-term should be investigated. Besides, more sensitive inflammatory outcome measures, such as cytokine profiling or central biomarkers based on neuroimaging or cerebrospinal fluid, could be used to further investigate the mechanistic interaction between glutamatergic modulation and immune pathology in BD.

Conclusion

In conclusion, adjunctive memantine showed no better results compared to placebo in alleviating the severity of manic symptoms, global clinical outcome, and altering peripheral inflammatory markers in patients with BD-I. Despite a strong biological rationale implicating NMDAR dysfunction and immune–glutamatergic interactions in the pathophysiology of BD-I, pharmacological NMDAR antagonism with memantine did not translate into clinically meaningful antimanic effects under rigorous randomized conditions. The large-scale, long-term investigations involving central inflammatory or neurobiological variables should be continued to provide more detailed information on whether particular subgroups can benefit through glutamatergic modulation.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

HM, RM, and AH contributed to developing the hypothesis and collecting data. EF performed the data analysis. SM, RM, and MN wrote the manuscript and designed the tables and figures. SR and HM contributed to the hypothesis, engineered and edited the data collection process, handled correspondence, and revised and finalized the manuscript prior to submission.

Funding

This study received no funding or grant support from any agency in the public, commercial, or not-for-profit sectors.

Data availability

Not applicable.

Declarations

Ethics approval

The study protocol was approved by the Institutional Review Board and Ethics Committee of Shiraz University of Medical Sciences (Approval Code: IR.SUMS.MED.REC.1403.783). The study was conducted in accordance with the ethical principles of the Declaration of Helsinki and its later amendments. The trial was registered in the Iranian Registry of Clinical Trials (IRCT20250430065539N1). All participants provided written informed consent prior to enrollment.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.